Freeze catalog signature vectors (T-614)

core/catalog/canonical.go

@@ -10,12 +10,15 @@ import (
 	"unicode/utf8"
 )
 
-var integerJSONNumber = regexp.MustCompile(`^-?(0|[1-9][0-9]*)$`)
+var integerJSONNumber = regexp.MustCompile(`^(0|[1-9][0-9]*|-[1-9][0-9]*)$`)
 
 func parseRestrictedJSON(data []byte) (any, error) {
 	if !utf8.Valid(data) {
 		return nil, fmt.Errorf("%w: input is not valid UTF-8", ErrInvalidDocument)
 	}
+	if err := validateJSONStringSurrogates(data); err != nil {
+		return nil, err
+	}
 	decoder := json.NewDecoder(bytes.NewReader(data))
 	decoder.UseNumber()
 
@@ -33,6 +36,78 @@ func parseRestrictedJSON(data []byte) (any, error) {
 	return value, nil
 }
 
+func validateJSONStringSurrogates(data []byte) error {
+	for index := 0; index < len(data); index++ {
+		if data[index] != '"' {
+			continue
+		}
+		next, err := scanJSONStringSurrogates(data, index)
+		if err != nil {
+			return err
+		}
+		index = next - 1
+	}
+	return nil
+}
+
+func scanJSONStringSurrogates(data []byte, start int) (int, error) {
+	for index := start + 1; index < len(data); index++ {
+		switch data[index] {
+		case '"':
+			return index + 1, nil
+		case '\\':
+			if index+1 >= len(data) {
+				return 0, fmt.Errorf("%w: incomplete string escape", ErrInvalidDocument)
+			}
+			if data[index+1] != 'u' {
+				index++
+				continue
+			}
+			codeUnit, ok := decodeJSONHexCodeUnit(data, index+2)
+			if !ok {
+				return 0, fmt.Errorf("%w: invalid unicode escape", ErrInvalidDocument)
+			}
+			switch {
+			case codeUnit >= 0xd800 && codeUnit <= 0xdbff:
+				if index+7 >= len(data) || data[index+6] != '\\' || data[index+7] != 'u' {
+					return 0, fmt.Errorf("%w: high surrogate is not paired", ErrInvalidDocument)
+				}
+				lowSurrogate, ok := decodeJSONHexCodeUnit(data, index+8)
+				if !ok || lowSurrogate < 0xdc00 || lowSurrogate > 0xdfff {
+					return 0, fmt.Errorf("%w: high surrogate is not followed by a low surrogate", ErrInvalidDocument)
+				}
+				index += 11
+			case codeUnit >= 0xdc00 && codeUnit <= 0xdfff:
+				return 0, fmt.Errorf("%w: low surrogate has no high surrogate", ErrInvalidDocument)
+			default:
+				index += 5
+			}
+		}
+	}
+	return 0, fmt.Errorf("%w: unterminated string", ErrInvalidDocument)
+}
+
+func decodeJSONHexCodeUnit(data []byte, start int) (uint16, bool) {
+	if start+4 > len(data) {
+		return 0, false
+	}
+	var value uint16
+	for _, digit := range data[start : start+4] {
+		value <<= 4
+		switch {
+		case digit >= '0' && digit <= '9':
+			value |= uint16(digit - '0')
+		case digit >= 'a' && digit <= 'f':
+			value |= uint16(digit-'a') + 10
+		case digit >= 'A' && digit <= 'F':
+			value |= uint16(digit-'A') + 10
+		default:
+			return 0, false
+		}
+	}
+	return value, true
+}
+
 func decodeJSONValue(decoder *json.Decoder) (any, error) {
 	token, err := decoder.Token()
 	if err != nil {